Abstract
A consensus forecast commonly represents a simple average, or equal weighting, of individual analyst forecasts. We demonstrate, however, that an equal-weight consensus generally does not aggregate information in individual analyst forecasts efficiently and that an aggregation scheme informed by forecast and analyst attributes results in a more accurate consensus forecast. For example, in the case of forecasts issued in the sixty days immediately preceding an earnings announcement, the mean squared error (MSE) of an informed consensus forecast whose component weights are based on forecast age, broker size, and analyst experience is significantly less than that of a naive consensus. In this case, MSE-improvement ranges from three to fifteen percent, respectively, as the number of forecasts in the consensus varies from four to twenty.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have